Skip to main content
The Polyvia API provides programmatic access to visual search and reasoning capabilities. Build applications that extract insights from charts, slides, and diagrams at scale.

​
Overview

The API is built around two core endpoints:

/ingest

Upload and index documents into your visual knowledge base

/query

Search and reason over your indexed visual content

​
Quick Example

Here’s a complete workflow: upload a set of financial reports, then query for insights.

​
1. Ingest Documents

import requests

api_key = "YOUR_API_KEY"
headers = {"Authorization": f"Bearer {api_key}"}

# Upload multiple documents
documents = [
    "apple_10k_2024.pdf",
    "apple_10k_2023.pdf",
    "microsoft_10k_2024.pdf",
    "microsoft_10k_2023.pdf",
    "google_10k_2024.pdf",
    "amazon_10k_2024.pdf"
]

for doc in documents:
    with open(doc, "rb") as f:
        response = requests.post(
            "https://api.polyvia.ai/ingest",
            headers=headers,
            files={"file": f},
            data={"name": doc}
        )
        print(f"Ingested {doc}: {response.json()['document_id']}")

​
2. Query Your Knowledge Base

# Query with visual reasoning
response = requests.post(
    "https://api.polyvia.ai/query",
    headers={**headers, "Content-Type": "application/json"},
    json={
        "query": "What's the difference between Apple and Microsoft revenue in Q4 2024?",
        "include_citations": True
    }
)

result = response.json()
print(result["answer"])
# "Apple Q4 2024 revenue: $119.58B [10-K page 24, chart]
#  Microsoft Q4 2024 revenue: $62.02B [10-K page 31, table]
#  Difference: $57.56B in favor of Apple"

for citation in result["citations"]:
    print(f"  - {citation['document']}, page {citation['page']}")

​
Next Steps

API Reference

Full endpoint documentation with all parameters

Polyvia MCP Server

Connect Polyvia to Claude Desktop and Cursor